Skip to content
Engineering / AI

AI Developer - AI Systems Integration & Automation

RemoteFull-TimeFull-time with required overlap with US working hours

This is not a no-code or basic automation role. The work is production software development: writing code in Python or in JavaScript, TypeScript and Node, integrating REST APIs with real authentication, and building custom middleware and API services when off-the-shelf tools fall short. Experience with n8n, Make or Zapier is useful here, but on its own it is not enough for this role. The judgment being hired for is knowing when an LLM is the wrong tool. A great deal of this work is ordinary deterministic logic that should never go near a model, and someone who reaches for AI at every step is not the right fit. Expect to be asked where you would use a model, where you would not, and what you would build instead. The portfolio or GitHub link on the application form is optional. Most of the strongest people in this field have their best work sitting in private client repositories, so leaving it blank costs you nothing and will not be counted against you.

Build and deploy AI-powered systems that automate IT and service-desk workflows end to end: LLM applications, REST API integrations and custom middleware that you write and ship yourself. This is production software development, not no-code automation. Remote, full-time with required overlap with US working hours.

Apply Now

About this role

This is a hands-on AI development role for someone who thinks like a systems architect. You would work inside a managed IT services environment, connecting ticketing systems, directory services and communication platforms into unified AI-driven workflows, and you would do it largely on your own: assessing the business problem, deciding what should be built, choosing the technical approach, and then writing and deploying the solution yourself. A major deliverable is an AI ticket-analysis system that reviews high volumes of daily support tickets and flags problems such as poor technician communication, missing approvals, incomplete documentation and non-standard actions, with explainable output and a feedback loop so it improves over time. Automation platforms have their place here, but they are not the job: where off-the-shelf tools fall short you would write the middleware and the production services yourself. Expect a technical interview built around real-world IT workflow architecture design, covering API selection, data flow, AI component design, authentication, error handling and deployment strategy.

What you’ll do

  • Design, build and deploy AI-powered systems that automate manual IT and service-desk processes end to end
  • Develop LLM-powered applications using OpenAI, Anthropic Claude, Google Gemini or equivalent APIs
  • Build robust REST API integrations with proper authentication, error handling and retry logic
  • Write custom middleware and production API services where off-the-shelf automation tools fall short
  • Automate multi-step IT service workflows such as user onboarding and offboarding and multi-step approval processes
  • Build an AI ticket-analysis system that reviews high volumes of daily support tickets and flags poor technician communication, missing approvals, incomplete documentation and non-standard actions
  • Provide explainable outputs and a feedback loop so the flagging keeps improving
  • Assess business problems end to end and recommend the correct technical architecture before building
  • Connect platforms such as ticketing systems, directory services and communication tools into unified AI-driven workflows
  • Own what you ship after launch: logging, monitoring, rate limiting and data validation

What we’re looking for

  • AI application development using LLMs (OpenAI, Anthropic Claude, Google Gemini or equivalent)
  • Prompt engineering and structured AI outputs
  • AI agent and workflow development
  • Production-level Python and/or JavaScript, TypeScript and Node.js
  • REST API integration: authentication, OAuth, API keys and JSON
  • Backend software development including databases, cloud services and deployment
  • Git and GitHub, and logging
  • Systems integration across multiple third-party platforms
  • Understanding of IT and managed-service workflows: ticketing, user onboarding and offboarding, approvals and access management
  • Custom middleware and API service development
  • Error handling, retries, rate limiting and data validation in integrations
  • Solution architecture and independent problem-solving
  • Mid to senior level: demonstrated delivery of real AI-powered applications and multi-system integrations in production environments

Nice to have

  • RAG, embeddings, vector databases and function or tool calling with LLMs
  • AI evaluation, hallucination control and reliability engineering
  • Experience with Autotask or a similar PSA or ticketing platform
  • Experience with NAVM or another managed-service integration platform
  • Active Directory and user permissions management
  • Automation tools such as n8n, Make, Zapier or Power Automate
  • Experience in a managed service provider, IT support or help-desk environment
  • SLA management and ticket escalation workflows
  • Internal dashboard and data pipeline development
  • SaaS or enterprise automation product development background

About F5 Global Talent

F5 Global Talent is a people-first company that hires experienced professionals and places them in long-term, fully remote roles with leading U.S. companies. We are your direct employer of record — we handle payroll, equipment, HR, and ongoing support, so you can focus on doing your best work. To date we have placed 500+ professionals, and 98% remain with their role past the first 90 days.

Apply for this role

Full-time, dedicated, remote. Complete the five steps below - it takes about fifteen minutes.

This is not a no-code or basic automation role. The work is production software development: writing code in Python or in JavaScript, TypeScript and Node, integrating REST APIs with real authentication, and building custom middleware and API services when off-the-shelf tools fall short. Experience with n8n, Make or Zapier is useful here, but on its own it is not enough for this role. The judgment being hired for is knowing when an LLM is the wrong tool. A great deal of this work is ordinary deterministic logic that should never go near a model, and someone who reaches for AI at every step is not the right fit. Expect to be asked where you would use a model, where you would not, and what you would build instead. The portfolio or GitHub link on the application form is optional. Most of the strongest people in this field have their best work sitting in private client repositories, so leaving it blank costs you nothing and will not be counted against you.

Step 1 of 5 — 20% complete

About you